// //

Saturday, December 3, 2016 - Welcome, guest user!

Great ride with Mike at Paisley; New Strava segment matching algorithm.

Monday, July 8, 2013 by  
Filed under Daily Blog

July
8
2013
My Fuel after yesterday's ride at the Paisley Woods MTB Trails

My Fuel after yesterday’s ride at the Paisley Woods MTB Trails

There were various rumblings on Facebook of a group mountain bike ride coming together at the Paisley Woods MTB Trails on Sunday morning, so Mike and I drove out there ready to ride. Apart from a local walking his dog, the trailhead was completely deserted. Oh well, sure can’t blame anyone for wanting to sleep in on a Sunday!

It was hot and very muggy, and after just a few miles we were both drenched in sweat–and this was at 8:00 AM! It was so humid that it felt like I was breathing more water than oxygen.

The trails at Paisley are very sandy in spots, but the recent rains have packed down a lot of the sand. So the trails are running fast, and Mike and I each set new PRs on the main 18.6 mile loop with spirited performances.

My average heart rate over the 18.6 mile main loop was 176 BPM. That. Freaking. Hurt!

Here’s my entire ride on Strava.

This minor 0.2 mile off-course error is why we believe Mike's segment did not match.

This minor 0.2 mile off-course error is why we believe Mike’s segment did not match.

Mike and I separated after just a couple miles, and so we were riding solo most of the time. The interesting thing about that is my ride matched the 18.6 mile “Paisley Clockwise” segment (the main loop) on Strava just fine, but Mike’s ride did not match. Late in the ride Mike made a wrong turn and traveled a mere 0.2 miles on a side trail (I think it was a hiking trail), before realizing the error and backtracking.

So, with the backtrack that was only a 0.4 mile diversion from a 18.6 mile segment. In other words, Mike rode a ~2.2% deviation from the recorded 18.6 mile segment, and because of that the segment did not match.

In the past Mike’s segment would have matched, but Strava has rolled out a new segment matching algorithm, and it is clearly far more strict than the old one. This is how Strava used to match segments with the old algorithm:

When we determine that a segment matches your activity, we look first at the start and endpoints, and make sure your data passes through those zones. Second, we make sure that at least 75% of the data in between matches the segment data. If any of these cases come up false, the segment will not match your activity or appear on your activity.

Mike had a good GPS lock with his Garmin Edge 500 yesterday, and his start and end points were dead-on. Looking at his recorded route, that was also dead-on the whole way (apart from the minor navigation error, of course). Even with the 0.4 mile navigation error (out and back) and accounting for some amount of normal GPS drift, Mike’s segment would have easily matched using the old method since a minimum of 95% of his recorded data matched the segment.

I don’t think Strava has publicly released any information on their new segment matching algorithm (I can find nothing in their knowledge base), but the new system seems much less tolerant of deviations (accidental or otherwise) from the recorded segment. We’ve noticed this behavior recently on various other segments, too.

Mike opened a ticket with Strava support, and perhaps they will be able to provide a bit more information on the new algorithm, and why the segment did not match. I’ll pass on any salient information from Strava in a future blog.

Speak Your Mind

Tell us what you're thinking...

You must be logged in to post a comment. Not yet a member? Registration is fast and free!